> > + if (!hidg->in_ep->enabled) {
> > + ERROR(hidg->func.config->cdev, "in_ep is disabled\n");
> > + status = -ESHUTDOWN;
> > + goto release_write_pending;
>
> What prevents "enabled" from changing right after you test this? You do
> not have the lock held so could it change then?
Inside usb_ep_queue() a similar check is used without blocking.
I suppose this is how it should be. Here it's necessary for the error message.
> > + }
> > +
> > status = usb_ep_queue(hidg->in_ep, req, GFP_ATOMIC);
> > - if (status < 0) {
> > - ERROR(hidg->func.config->cdev,
> > - "usb_ep_queue error on int endpoint %zd\n", status);
> > + if (status < 0)
> > goto release_write_pending;
>
> No error message if queueing up failed?
Yes. For each write from userspace, this generates one message.
This can lead to a huge spam in the log.
